In the clinical evaluation of left ventricular (LV) function in humans, there has been growing interest in the diagnostic and prognostic power of measurements or derived indexes of LV systolic relaxation, ie, LV isovolumetric pressure fall and early rapid filling. Most of these measurements have been appreciated as excellent predictors of LV systolic function, in particular in the early phases of ischemic and hypertrophic cardiomyopathies. Misinterpretation of such measurements, however, has often contributed to the many controversies concerning the diagnosis of diastolic cardiac failure. To most clinicians, the concept and diagnosis of LV diastolic dysfunction or failure are not easy. Part of this problem relates to the fact that many, mainly clinical, investigators erroneously persist in applying different definitions of diastole, mostly on empirical-historical grounds, depending on whether they consider the heart as a pump instead of a “muscular” pump or analyze cardiac hemodynamics as a function of time or as pressure-volume (P-V) relations. Meanwhile, diastolic failure of the heart has become a widely recognized clinical entity. In a recent review, we identified diastolic failure as a condition resulting from an increased resistance to ventricular filling and leading to symptoms of congestion due to an inappropriate shift of the diastolic P-V relation.1 Causes of diastolic failure, and hence of shifts in the diastolic P-V relation, are inappropriate tachycardia, decreased diastolic myocardial/ventricular compliance, and impaired LV systolic relaxation, ie, impaired LV isovolumetric pressure fall or early rapid filling.
Further, in this issue of Circulation, Ishizaka and colleagues2 have again drawn attention to the importance of impaired LV systolic relaxation as an early event in experimental heart failure. These investigators examined the contribution of changes in end-systolic volume and of the loading sequence to impaired LV isovolumetric relaxation in overdrive-pacing heart failure. They used caval occlusion to decrease end-diastolic volume and quantified LV isovolumetric relaxation by the time constant tau. They found larger changes in tau for a comparable reduction in end-systolic force in heart failure compared with the normal heart. The authors discuss possible mechanisms by which systolic loading may differently alter LV isovolumetric relaxation in normal and in failing hearts: restoring forces, intracellular calcium handling, additional crossbridge recruitment, nonuniformity, and wave reflection. In isolated cardiac muscle, restoring forces are known to affect primarily isotonic relaxation, ie, muscle fiber lengthening. Because in the ventricle, a substantial portion of fiber lengthening already occurs during LV isovolumetric relaxation (untwisting and shape changes), restoring forces will contribute not only to early rapid ventricular filling but also, to a large extent, to isovolumetric pressure decline. In view of the widely recognized importance of impaired LV systolic relaxation in the early phases of heart failure or as a possible cause of diastolic failure as defined above, these interesting observations can be fully appreciated only when interpreted in the somewhat broader context of the heart as a muscular pump.
Load Dependence of Relaxation
Both in isolated cardiac muscle and in the intact ventricle, many investigators have emphasized the conceptual link between the processes underlying activation during contraction and those underlying inactivation during relaxation, relaxation being governed by the interaction of detachment of cross-bridges and Ca2+ reuptake by the sarcoplasmic reticulum and modulated by prevailing load; we called this latter property the “load dependence” of relaxation.3 4 5 As a consequence of load dependence, rapid isotonic lengthening during relaxation of a preloaded or afterloaded isotonic twitch in isolated cardiac muscle contrasts with the slower force decline of either an isometric twitch or an afterloaded isotonic twitch regardless of whether isometric force decline preceded or followed lengthening.
To understand the concept of load dependence, one should keep in mind the inactivation processes that underlie isotonic lengthening as well as isometric force decline during relaxation. Isotonic and isometric relaxation are governed by the ensemble of processes leading to the disappearance of force-generating cross-bridges. The number of cross-bridges during relaxation is determined by (1) the life cycle of each individual cross-bridge along with regulatory properties of the contractile proteins and (2) calcium removal by the calcium-sequestering membrane systems, particularly the sarcoplasmic reticulum. Although both isotonic and isometric relaxation modes are governed by the same determinants of cross-bridge kinetics, the relative contributions of these determinants in controlling onset and rate of relaxation are different in isometric force decline and in isotonic lengthening. In an isometric twitch, “cooperative” activity, that is, a process of “increased” sensitivity of the contractile proteins that is induced by the attachment of cross-bridges and hence by the force development itself, will upgrade the development and maintenance of force throughout contraction and relaxation. On the other hand, calcium sequestration by the sarcoplasmic reticulum will, in the presence of a reduced effect of cooperative activity in the isotonic twitch and through facilitation of cross-bridge detachment, allow for load-induced rapid lengthening.
From these considerations, it was postulated that in isotonic lengthening, in the presence of a well-functioning Ca2+ reuptake by the sarcoplasmic reticulum, cross-bridges would detach more easily than in isometric force decline because of the additional effect of loading in isotonic conditions. Preloaded or afterloaded isotonic twitches, therefore, are always of shorter duration than the corresponding isometric twitch; load dependence of relaxation is thereby manifest as the resulting separation in time of relaxation in isotonic and isometric twitches or as the resulting difference in twitch duration. Hence, the extent of load dependence in isolated cardiac muscle can be derived from the pattern (eg, onset, peak rate, and rate versus time) of both isotonic lengthening and isometric force decline during twitch relaxation. Accordingly, these events are an integral part of one and the same contraction-relaxation cycle and hence of muscle “systole.” To cardiac muscle physiologists, muscle “diastole” refers to the rest or pause between two such activity cycles.5
Load Dependence In Vivo
Similarly, in the ventricle, isovolumetric pressure fall and early rapid filling somehow, although in an auxotonic fashion,5 relate to tension decline and rapid lengthening of the ventricular myocardium, respectively; ventricular relaxation should therefore, on conceptual grounds, be considered as part of systole. From this it follows that systolic function of the heart as a muscular pump is a reflection of the interaction between the loading conditions (hemodynamics, Laplace, twisting-untwisting) and the intrinsic properties of the myocardium during contraction (contractility) and relaxation (load dependence) and is modulated by some degree of nonuniformity in space and time.
Interestingly, the concept of load dependence of ventricular relaxation has recently been endorsed by in vivo canine experiments. As stated above, in the ventricle a substantial portion of cardiac fiber lengthening already occurs during isovolumetric pressure fall. Load dependence in vivo may therefore become manifest not merely as load-induced rate changes during early rapid filling but during isovolumetric pressure fall as well, ie, tau, dP/dt(−), etc. In an attempt to answer the question of which ejection variables are important in determining the increase in isovolumetric LV relaxation rate with increased stroke volume, Hori and colleagues6 used an isolated canine heart with servocontrolled volume; this model is particularly suited to control the entire pattern of LV volume during systole and diastole and to isolate the most critical ejection parameters. The authors demonstrated that the increased rate of isovolumetric relaxation, measured as the time constant tau of isovolumetric LV pressure fall, was a mere consequence of the delayed onset of relaxation induced by increased stroke volume. However, the influence of pressure variables, such as end-systolic force or stress, was not directly addressed in this article. Moreover, the time constant of isovolumetric LV pressure decline does, unfortunately, merely characterize a limited portion of ventricular relaxation, ie, from time to minimal dP/dt to 5 or 10 mm Hg above minimal LV pressure. Leite-Moreira and Gillebert7 subdivided LV pressure decline during relaxation into three phases and studied the entire pattern of LV pressure decline by pressure phase-plane analysis, following up the force phase-plane analyses in our earlier article.8 By taking into account measured peak isovolumetric LV pressure of a control beat, they could predict the effect of abrupt elevations in systolic LV pressure on the rate of LV pressure decline during the three phases. The J-shaped curve relating time constant tau to LV systolic pressure, in their Figs 7 and 8, reflects the effects of loading on ventricular relaxation rate in the presence of unaltered intrinsic load dependence. Altered load dependence would become manifest as a shift in this curve (Fig 1A⇓). Clearly, this kind of analysis must be taken into account whenever ventricular rate is examined as a measurement to evaluate late systolic events during LV relaxation and filling, in particular when loading conditions are expected to be affected, as, for example, after pharmacological interventions or in clinical situations. Both these in vivo studies illustrate that ventricular relaxation dynamics of normal hearts are, under well-controlled neurohumoral conditions, uniquely determined by the loading conditions (hemodynamics, Laplace, untwisting) interacting with intrinsic load dependence of relaxation5 ; as outlined above, this latter property is determined by Ca2+ reuptake and affinity of the contractile proteins.
In the study by Ishizaka and colleagues,2 LV isovolumetric relaxation was compared before and after induction of heart failure in one set of seven conscious dogs. Heart failure, as induced by overdrive pacing, is known to be characterized by early functional disturbances, not only in contractile function and in endothelium- and β-adrenergic receptor–mediated control but also in relaxation. These functional disturbances persist for different time intervals after cessation of pacing.9 This may be consistent with different structural changes, such as in the myocardium,10 in the interstitium,11 and in the endocardium.12 In this model of congestive heart failure by overdrive pacing, the presence of relaxation abnormalities again draws our attention to the diagnostic significance of measurements or indexes during this phase.
The individual dog data in Fig 3 of Ishizaka et al2 do not exclude the possibility that all the data points both before and after pacing might still fall on one and the same J-shaped curve, ie, similar to Figs 7 and 8 in the paper by Leite-Moreira and Gillebert,7 and hence demonstrating unchanged load dependence in cardiac failure. The increased slope after pacing would then merely follow from a shift along the abscissa to higher end-systolic force on one and the same uninterrupted curve relating tau to end-systolic force in the failing hearts; in other words, the data could be interpreted by a mere increase in loading conditions (Fig 1B⇑). This would also explain the equalization of the changes in tau when stresses and heart rate were matched at the same levels in control and failing hearts.13 Referring to loading sequence as a determinant of isovolumetric relaxation rate, Ishizaka et al noticed that the observed changes in tau in the failing heart were related to time to peak force rather than to end-systolic volume, despite substantial residual variation in tau at any given time to peak force (Fig 6 in Reference 2). Heart failure did not result in a shift in the curve relating tau to loading sequence and hence, as outlined above, did not result in a change in intrinsic load dependence. To further solve the problem of whether impaired relaxation in pacing-induced heart failure is due to a mere change in prevailing loading conditions or sequence, as suggested by Ishizaka et al, or to a change in intrinsic load dependence of muscular relaxation, which would imply a change in the failing heart either in the Ca2+ reuptake or in the affinity of the contractile proteins, a more elaborate series of experiments has to be carried out encompassing the entire relaxation process at matched loading conditions, as in the study by Leite-Moreira and Gillebert.
Implications for the Evaluation of LV Systolic Relaxation
First, from these conceptual and experimental considerations at both the muscular and the ventricular levels, it remains clear that contraction and relaxation events together with the underlying activation-inactivation processes are part of one and the same activity cycle, hence systole. Diastole defines the events between two such activity cycles and can be characterized from diastolic P-V relations and compliance measurements. By contrast, all measurements or derived indexes related to ventricular relaxation, ie, to LV isovolumetric pressure fall and to early rapid filling, should be considered systolic. Identifying these late-systolic events during LV relaxation as diastolic is, in view of the above considerations of the heart as a muscular pump, not merely a semantic misnomer but rather conceptually inappropriate. Although impaired systolic relaxation may cause an inappropriate upward shift of the diastolic P-V relation and hence may lead to diastolic dysfunction or failure, in the absence of such shifts these late-systolic abnormalities should be considered as a sign of general, often early, systolic dysfunction.
Second, in evaluating these LV relaxation abnormalities, it is essential that one differentiate between prolonged contraction and impaired relaxation (Fig 2⇓). This distinction will follow from a close analysis of rate and timing of relaxation indexes. Prolonged contraction is due to a delayed onset of relaxation regardless of concomitant variations in rate; it is physiological, compensatory, and not by itself deleterious and does not normally shift the diastolic P-V relation, at least not at appropriate heart rates. Onset of relaxation can be modulated by varying systolic (pressure or volume) loading, by the cardiac endothelium, and by various drugs. By contrast, impaired systolic relaxation may induce an upward shift of the diastolic P-V relation, thus leading to diastolic failure as defined above; it is pathophysiological and characterized by a decreased rate or extent of pressure decline and rapid filling. Causes of impaired systolic relaxation include (1) diminished intrinsic load dependence due to impaired (in)activation (Ca2+ handling, sarcoplasmic reticulum function, contractile protein properties), (2) excessive changes in load, and (3) inappropriate nonuniformity of load and (in)activation in time and space. Ironically, since one would not hesitate to label as “systolic” LV relaxation events with predominant changes in timing (prolonged contraction), why then, for the same portion of the cardiac cycle, label as “diastolic” those with predominant changes in rate (impaired relaxation)?
Third, a full evaluation of systolic LV relaxation should, therefore, encompass at least three types of measurement: (1) relaxation rate (dP/dt[−], tau, isovolumetric duration, rapid filling rate, E wave, etc), (2) the time interval from the onset of systole to the instant at which these rate indexes were measured, and (3) the rate pattern of the entire relaxation process (initial/early, intermediate/middle, and terminal/late phases of relaxation). As a consequence, frequently used measurements such as peak dP/dt(−), time constant tau of a selected part of isovolumetric relaxation, or early peak filling rate open a too narrow window on the relaxation process. Erroneous conclusions might often be prevented by an appropriate analysis of timing and rate pattern of LV systolic relaxation.
The opinions expressed in this editorial are not necessarily those of the editor or of the American Heart Association.
- Copyright © 1995 by American Heart Association
Ishizaka S, Asanoi H, Wada O, Kameyama T, Inoue H. Loading sequence plays an important role in enhanced load-sensitivity of left ventricular relaxation in conscious dogs with tachycardia-induced cardiomyopathy. Circulation. 1995;92:3560-3567.
Brutsaert DL, Housmans PR, Goethals MA. Dual control of relaxation: its role in the ventricular function in the mammalian heart. Circ Res. 1980;47:637-652.
Brutsaert DL, Rademakers FE, Sys SU. Triple control of relaxation: implications for the cardiac patient. Circulation. 1984;69:190-196.
Brutsaert DL, Sys SU. Relaxation and diastole of the heart. Physiol Rev. 1989;69:1228-1315.
Hori M, Kitakaze M, Ishida Y, Fukunami M, Kitabatake A, Inoue M, Kamada T, Yue DT. Delayed end ejection increases isovolumic ventricular relaxation rate in isolated perfused canine hearts. Circ Res. 1991;68:300-308.
Leite-Moreira AF, Gillebert TC. Nonuniform course of left ventricular pressure fall and its regulation by load and contractile state. Circulation. 1994;90:2481-2491.
Sys SU, Brutsaert DL. Determinants of force decline during relaxation in isolated cardiac muscle. Am J Physiol. 1989;257:H1491-H1498.
Moe GW, Stopps TP, Howward RJ, Armstrong PW. Early recovery from heart failure: insight into the pathogenesis of experimental chronic pacing-induced heart failure. J Lab Clin Med. 1983;112:426-432.
Zellner JL, Spinale FG, Eble DM, Hewett KW, Crawford FA Jr. Alterations in myocyte shape and basement membrane attachment with tachycardia-induced heart failure. Circ Res. 1991;69:590-600.
Weber KT, Pick R, Silver MA, Moe GW, Janicki JS, Zucker IH, Armstrong PW. Fibrillar collagen and remodeling of dilated canine left ventricle. Circulation. 1990;82:1387-1401.
Andries LJ, Kaluza G, Sys SU, Brutsaert DL. Rapid ventricular pacing in rabbits alters the F-actin cytoskeleton in endocardial endothelial and subendothelial cells. Pflugers Arch. 1995;430(suppl 4):R66.
Komamura K, Shannon RP, Pasipoularides A, Ihara T, Lader AS, Patrick TA, Bishop SP, Vatner SF. Alterations in left ventricular diastolic function in conscious dogs with pacing-induced heart failure. J Clin Invest. 1992;89:1825-1838.